Hacker News Reader

Comment by auntienomen | Hacker News Reader

This seems like incredibly bad design. Is there any justification or use case where someone would want the abstraction to leak like this?

View on HN

https://en.wikipedia.org/wiki/Significant_figures

Trailing zeroes convey information about the measurement precision of a number. If you're doing math with real-world measurements, errors accumulate, and you're expected to keep track of them.

All true, but using a BigDecimal to do that is abusing the implementation and keeping important assumptions implicit. Better to explicitly quantify the uncertainty with a second variable.

If one mass is 1.0 kg ± 40 g, and another mass is 1.000 kg ± 0.4 g, are they close enough to equal? Maybe, and if you round them to the same precision you get whichever answer you wanted.

It is likely because equals is usually “object equals”, not necessarily “math equals”. Since the internal representation is different, the objects are indeed different. Tricky though, and most 3rd party code would likely choose to make equals true in this kind of situation.

Java has tons of examples where the core behavior is logically correct, but VERY unintuitive - I feel the pain daily.

I don't think that flies, because the whole point of the .equals() method in regards to string is to avoid "object equals" and make sure the strings say the same thing..

It's supposed to be a mess. We have very little job security anymore.

Eh just par for the course for Java https://news.ycombinator.com/item?id=21765788

Reply on news.ycombinator.com